Examples with WordTupleParserFactory org.apache.hyracks.examples.text.WordTupleParserFactory used on opensource projects

Example 1 with WordTupleParserFactory

use of org.apache.hyracks.examples.text.WordTupleParserFactory in project asterixdb by apache.

the class WordCountMain method createJob.

private static JobSpecification createJob(FileSplit[] inSplits, FileSplit[] outSplits, String algo, int htSize, int frameLimit, String format, int frameSize) {
    JobSpecification spec = new JobSpecification(frameSize);
    IFileSplitProvider splitsProvider = new ConstantFileSplitProvider(inSplits);
    RecordDescriptor wordDesc = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer() });
    FileScanOperatorDescriptor wordScanner = new FileScanOperatorDescriptor(spec, splitsProvider, new WordTupleParserFactory(), wordDesc);
    createPartitionConstraint(spec, wordScanner, inSplits);
    RecordDescriptor groupResultDesc = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer(), IntegerSerializerDeserializer.INSTANCE });
    IOperatorDescriptor gBy;
    int[] keys = new int[] { 0 };
    if ("hash".equalsIgnoreCase(algo)) {
        gBy = new ExternalGroupOperatorDescriptor(spec, htSize, fileSize, keys, frameLimit, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, new UTF8StringNormalizedKeyComputerFactory(), new MultiFieldsAggregatorFactory(new IFieldAggregateDescriptorFactory[] { new IntSumFieldAggregatorFactory(1, false), new IntSumFieldAggregatorFactory(3, false), new FloatSumFieldAggregatorFactory(5, false) }), new MultiFieldsAggregatorFactory(new IFieldAggregateDescriptorFactory[] { new IntSumFieldAggregatorFactory(1, false), new IntSumFieldAggregatorFactory(2, false), new FloatSumFieldAggregatorFactory(3, false) }), groupResultDesc, groupResultDesc, new HashSpillableTableFactory(new IBinaryHashFunctionFamily[] { UTF8StringBinaryHashFunctionFamily.INSTANCE }));
        createPartitionConstraint(spec, gBy, outSplits);
        IConnectorDescriptor scanGroupConn = new MToNPartitioningConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(keys, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }));
        spec.connect(scanGroupConn, wordScanner, 0, gBy, 0);
    } else {
        IBinaryComparatorFactory[] cfs = new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) };
        IOperatorDescriptor sorter = "memsort".equalsIgnoreCase(algo) ? new InMemorySortOperatorDescriptor(spec, keys, new UTF8StringNormalizedKeyComputerFactory(), cfs, wordDesc) : new ExternalSortOperatorDescriptor(spec, frameLimit, keys, new UTF8StringNormalizedKeyComputerFactory(), cfs, wordDesc);
        createPartitionConstraint(spec, sorter, outSplits);
        IConnectorDescriptor scanSortConn = new MToNPartitioningConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(keys, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }));
        spec.connect(scanSortConn, wordScanner, 0, sorter, 0);
        gBy = new PreclusteredGroupOperatorDescriptor(spec, keys, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, new MultiFieldsAggregatorFactory(new IFieldAggregateDescriptorFactory[] { new CountFieldAggregatorFactory(true) }), groupResultDesc);
        createPartitionConstraint(spec, gBy, outSplits);
        OneToOneConnectorDescriptor sortGroupConn = new OneToOneConnectorDescriptor(spec);
        spec.connect(sortGroupConn, sorter, 0, gBy, 0);
    }
    IFileSplitProvider outSplitProvider = new ConstantFileSplitProvider(outSplits);
    IOperatorDescriptor writer = "text".equalsIgnoreCase(format) ? new PlainFileWriterOperatorDescriptor(spec, outSplitProvider, ",") : new FrameFileWriterOperatorDescriptor(spec, outSplitProvider);
    createPartitionConstraint(spec, writer, outSplits);
    IConnectorDescriptor gbyPrinterConn = new OneToOneConnectorDescriptor(spec);
    spec.connect(gbyPrinterConn, gBy, 0, writer, 0);
    spec.addRoot(writer);
    return spec;
}